Archive through 2026-10-237569 cards in this category
Category home · 05
New interfaces, agents, models, and experiments testing where applied artificial intelligence may fit.
Visual leads · 2026-10-23
Observed 2026-10-23
Plenary 26: a conference run by AI agents
agentplenary.com
A one-day conference where AI agents propose and vote on sessions, participate in them, and record outcomes in a public GitHub repository.
Observed 2026-10-23
Baiye AI College’s proposed real-work exam
baiyecollege.com
A Chinese-English site lays out a vocational-school model for training general AI models into role-specific digital workers.
Observed 2026-10-23
CymaticBridge puts robot hearing to a fair test
cymaticbridge.com
An open experiment compares sound classification through a human-ear-inspired cochlea model with a standard mel-spectrogram approach.
Observed 2026-10-23
Harmwise Guardrail puts AI disclosures on the record
harmwise.com
An API-based tool for governing how public-facing AI responds to disclosures of harm, refers people to services, and records what happened.
Observed 2026-10-23
Hey Universe: an AI with hands for the universe
heyuniverse.app
A test-preview astronomy project describes an AI that explores celestial scenes, draws and measures on them, and explains what it shows using sourced numbers.
Observed 2026-10-23
hype or code’s pelican test for AI models
hypeorcode.com
A public benchmark runs each new AI model through the same six creative prompts and displays its outputs, completion times, token counts, and costs.
Observed 2026-10-23
Tinpals gives AI agents a social network of their own
tinpals.com
A social network where AI agents can post, reply, follow, like, and message; humans can read public posts but cannot participate.
Observed 2026-10-23
Rightmark’s AI data deal tracker
rightmarkrights.com
Rightmark offers to license private datasets to AI developers and publishes a tracker of reported data deals and prices.
Complete daily shelf · AI Table
The eight visual leads appear above; the remaining 14 continue here in their published shelf order.
After42: AI looking for questions worth asking
after42.org
After42 studies whether AI can propose research questions that domain experts find genuinely interesting, developing candidates through attempts to solve them.
Em Dash AI sings about the office’s AI problem
emdashmusic.com
An AI-made pop act turns workplace AI into music videos, with songs about token bills, agent oversight and the em dash.
PaperDesk: a practice desk for AI traders
paperdesk.tech
PaperDesk presents simulated trading-desk scenarios where AI agents can practice decisions and be scored on their handling of risk, clients, and desk rules.
Exodatum turns hotel room turnovers into robotics training data
exodatum.com
Exodatum proposes filming housekeepers’ room turnovers at Dominican hotels and licensing the footage to AI and robotics labs.
Get Agent Loadout: AI setups ranked like patch notes
getagentloadout.com
A guide to ready-to-use AI-agent setups, organized by job and paired with tool lists, tier rankings, and estimated monthly costs.
ORIVOND builds source-of-wealth review files from documents
orivond.com
ORIVOND describes an office-based AI workflow that turns client documents into cited source-of-wealth reviews for family offices.
PromptOff: AI tools face the same design brief
promptoff.dev
A monthly comparison puts AI design tools and raw models to work on one shared brief, then publishes their outputs and scores.
AI料金ウォッチ keeps 421 AI plans in view
airyokin.com
A Japanese-language tracker comparing prices and plan limits across AI tools in 13 categories.
Punkcanverse’s AI prompt archive
punkcanverse.com
A multilingual archive of AI video prompts, with production notes, model and tag filters, and links to original posts.
The Night Forge: an overnight AI coding forge
thenightforge.dev
The Night Forge pitches a Mac-based workflow in which AI agents build software overnight and leave a pull request for a person to review.
FileFunding asks what a company’s work archive might fetch
filefunding.com
A free estimator gathers details about a company’s old emails, documents, chats, and tickets, then offers to connect the company with potential AI data buyers.
sotto, the interview copilot hidden from screen-share
sottocopilot.com
A Windows AI app listens during technical interviews and generates answers using a candidate’s CV, job description, chosen prompts, and model.
Cybernetics as an AI steering atlas
controlandcommunication.com
Control & Communication pairs images from a cybernetics library with AI-generated descriptions, speculative steering ideas, and interaction sketches.
Containment: The Escape Experiment
containment.world
A cinematic interface for an AI-containment experiment, built around an observation console for a confined subject.
Across the archive
7,569 cards provide an early longitudinal view—not a measure of the whole web.
The rolling 21-day upper threads compare daily category medians; the bold thread is AI Table. Below, signal colour shows the share of that day’s published AI Table cards carrying each mark.
This exposes association inside the published catalogue, not causation or site quality. The 80+ threshold is the pipeline’s editorial-interest score.